Predictably Irrational – A Book Review

Posted on June 7, 2008 by


Back in January I posted a review of the latest Tim Harford book – The Logic of Life.  

Here is how I described this book:

“The Logic of Life” can be described as a book that takes the standard assumption of rationality in economics for a ride.  In essence, Harford examines a whole collection of diverse stories – most of these outside the standard topics covered in economics – by assuming first that people are making choices rationally.

Much of standard economics rests upon the assumption of rationality.  And although human beings are frequently “rational”, behavioral economists seem to delight in finding examples where this is not the case.

And now we have an entire book that offers a collection of such examples.  Predictably Irrational: The Hidden Forces that Shape Our Decisions– by Dan Ariely – provides story after story illustrating just how irrational the human animal can be.

What We Learn from Experiments

Before we get to some of these stories, we must first note the methodology followed by Ariely (and the co-authors on his journal publications).  Economists typically investigate actual human behavior via regression analysis.  For example, in The Wages of Wins we argue that coaches and general managers in the NBA over-value scoring (and under-value shooting efficiency, turnovers, etc…).  This conclusion is based on regressions examining the determinants of free agent salary and the coaches’ voting for the All-Rookie team. 

Ariely’s research does not utilize the standard regression approach, but rather employs controlled experimentation.  Basically Ariely and his co-researchers constructed experiments to see how people behave and make decisions.  And this research indicated that people are not quite the rational decision-makers posited in standard economic theory.

For example, consider the standard supply and demand model (which I teach every year).  Ariely’s experiments demonstrated that the consumer’s willingness to pay – which is supposed to represent the independent preferences of consumers – can be easily manipulated.  Specifically, because people can be “anchored” to initial assessments – even if those initial values are arbitrary – manufacturers can influence how people value an item.

Beyond the supply and demand issue, Ariely’s research also examines

  • how owning an item increases our value of the item.
  • how having more choices makes us worse off.
  • how prices impact how much utility we get from an item.
  • the problem of procrastination.
  • why we are dishonest and how such behavior can be limited.
  • how people’s behavior can worsen when you move from social norms (you should not do something because it is frowned upon by your fellow human beings) to market norms (you should not do something because it will cost you money).
  • how our sexual attitudes depend upon our state of arousal (a rather interesting experiment).

The Value of Blind Peer Review

Although all of these stories were excellent, one particular experiment really captured my attention.   In an experiment with students and beer, Ariely was able to demonstrate how people’s expectations impacted their final assessment.  Why is this important?

Ariely argued that people tend to get what they expect.  Consequently it’s very hard for us to evaluate something objectively if the evaluation begins with pre-conceived notions.  This means it’s hard for Republicans to hear the arguments of Democrats – and Democrats to hear the arguments of Republicans – because once a person knows the argument is coming from a person of the opposite party the argument is immediately discounted.

Although this is important for politics, this facet of human behavior also impacts how we evaluate academic research. 

In academia we prefer to see research evaluated by blind peer review. What does this mean? The “peer” issue means that research has to be evaluated by people who understand the material being presented.  So we don’t have economists evaluating the work of chemists (or chemists evaluating the work of economists).

What about the “blind” issue?  Again, Ariely’s experiments demonstrate that expectations impact evaluations.  For research, this means that who wrote a paper will influence how we see the paper.  Because academics know this effect exists, we tend to insist that papers be reviewed blindly.  This means the reviewer must be a person who does not know who wrote the paper and doesn’t have a vested interest in how the paper is judged.

Now this ideal is not always achieved.  Frankly, you can sometimes tell who wrote a paper.  And even if you can’t tell exactly who, the author or authors of a paper often reveal something about themselves that can influence your judgment.  For example, you can almost always tell when a paper is written by a graduate student or an inexperienced researcher. And yes, that fact will impact your evaluation.  Still, blind review – despite its problems – is better than the alternative.

The alternative we see played out on-line all the time.  Research posted on-line is not subject to blind review.  In fact, it’s often not subject to peer review (for example, if you have no training in econometrics you are going to have trouble reviewing a paper relying on econometric methods to tell its story).  As a consequence, the quality of the reviewer’s comments on-line – relative to what I see from academic journals – is quite poor. 

For the most part one can predict an on-line reaction if you simply know

1. the relationship between the reviewer and the author of the study.  Friends do not criticize friends on-line.  On the other hand, enemies do nothing but criticize.

2. related to #1 is the reviewer’s beliefs relative to the conclusions of the on-line study.  Many on-line people seem to argue from conclusion back to evidence (as opposed to from evidence to conclusion).  Obviously if you start with the conclusion, you are not going to like the study if you don’t like what it concludes.

My sense is that Ariely’s research – and my defense of the blind/peer review process – will not convince people who present and review research on-line.  And that’s also not surprising.

At least, not if you have read Super Crunchers by Ian Ayres. My hope is to review this book in the next few weeks. In the meantime, I encourage everyone to read Predictably Irrational.  And then Super Crunchers.  Perhaps this could be the beginning of the WoW Journal Book Club.

– DJ